Re: [sched/balance] INFO: possible recursive locking detected

[Fengguang Wu <fengguang.wu@intel.com>]

On Thu, Mar 13, 2014 at 03:43:11PM +0800, Alex Shi wrote:
> On 03/03/2014 10:25 PM, Fengguang Wu wrote:
> > Greetings,
> > 
> > I got the below dmesg and the first bad commit is
> > 
> > https://github.com/alexshi/power-scheduling.git single-balance.bak
> > commit 2adc591a5cb6768ad11e63be9011906aa171148d
> > Author:     Alex Shi <alex.shi@linaro.org>
> > AuthorDate: Fri Feb 28 15:25:09 2014 +0800
> > Commit:     Alex Shi <alex.shi@linaro.org>
> > CommitDate: Sun Mar 2 09:16:38 2014 +0800
> > 
> >     sched/balance: replace new_idle_balance
> >     
> >     Curren idle_balance will pull task on half of this group. That is
> >     not needed in central balance.
> 
> 
> Thanks for your testing, Fengguang!
> 
> I fixed this issue and updated to the git tree. Is the single-balance
> branch still in your performance radar? :)

Sure, however you know it's costly to queue tests for one single tree.
If your trees are directly based on -rcX kernels, that would cut the
tests required for comparison by half.

Thanks,
Fengguang